Interfollicular Fibrosis in the Thyroid of the Harbour Porpoise: An... Disruption?

Interfollicular Fibrosis in the Thyroid of the Harbour Porpoise: An Endocrine
Disruption?
Krishna Das,
1,6
Arndt Vossen,
1
Kristal Tolley,
2
Gisli Vkingsson,
3
Kristina Thron,
4
Gundi Mller,
5
Wolfgang Baumgrtner,
5
Ursula Siebert
1
1
Research- and Technology Centre Westcoast, Christian-Albrechts-University Kiel, Bsum, Germany
2
Leslie Hill Molecular Systematics Laboratory, Kirstenbosch Research Centre, South African National Biodiversity Institute, Private Bag X7, Claremont
7735, South Africa
3
Marine Research Institute, Reykjavk, Iceland
4
Baltic Sea Research Institute Warnemuende, Warnemuende, Germany
5
School of Veterinary Medicine Hannover, Hannover, Germany
6
Mare Center, Laboratory for Oceanology B6c, Lige University, Lige, Belgium
Received: 21 April 2005 /Accepted: 19 September 2005
Abstract. Previous studies have described high levels of
polychlorobiphenyls (PCB), polybrominated diphenylether
(PBDE), toxaphene, p,p0-dichlorodiphenyltrichloroethane
(DDT), and p,p0-dichlorodiphenyldichloroethylene (DDE) in
the blubber of the harbour porpoise from the North Sea raising
the question of a potential endocrine disruption in this species.
In the present study, the thyroids of 57 harbour porpoises from
the German and Danish (North and Baltic Seas), Norwegian,
and Icelandic coasts have been collected for histological and
immunohistological investigations. The number of follicles and
the relative distribution of follicles, connective, and solid tis-
sues (%) were quantified in the thyroid of each individual.
Then, the potential relationship between the thyroid mor-
phometry data and previously described organic compounds
(namely, PCB, PBDE, toxaphene, DDT, and DDE) was
investigated using factor analysis and multiple regressions.
Thyroid morphology differed strongly between sampling sites.
Porpoises from the German (North and Baltic Seas) and Nor-
wegian coasts displayed a high percentage of connective tissues
between 30 and 38%revealing severe interfollicular fibrosis
and a high number of large follicles (diameter >200 lm). A
correlation-based principal component analysis (PCA) revealed
two principal components explaining 85.9%of the total vari-
ance. The variables PCB, PBDE, DDT, and DDE compounds
loaded highest on PC1 whereas toxaphene compound loaded
most on PC2. Our results pointed out a relationship between
PC1 (PCBs, PBDE, DDE, and DDT compounds) and interfol-
licular fibrosis in the harbour porpoise thyroids. Such an
association is not alone sufficient for a cause–effect relationship
but supports the hypothesis of a contaminant-induced thyroid
fibrosis in harbour porpoises raising the question of the long-
term viability in highly polluted areas.
The harbour porpoise (Phocoena phocoena) is the most com-
mon cetacean species in the North and Baltic Seas (Hammond
et al. 2002). Beside substantial incidental catches, competition
with fishery, noise disturbance, and chemical pollutants are
likely to contribute to its decrease in recent years (Benke et al.
1998; Hammond et al. 2002; Read 2002).
The harbour porpoise as a predator may accumulate high
concentrations of persistent inorganic and organic pollutants in
its tissue (Aguilar and Borrell 1995; Das et al. 2003; Reijnders
and Aguilar 2002). Organohalogens are industrial chemicals,
by-products of the chemical industry, or pesticides, and may
exert toxic effects on wild and laboratory organisms (Wright
and Welbourn 2002). Organohalogens such as toxaphene,
polybrominated diphenylether (PBDE), polychlorinated bi-
phenyls (PCB), dichlorodiphenyltrichloroethane (DDT), and
dichlorodiphenyldichloroethylene (DDE) compounds are
strongly suspected to have endocrine-disrupting effects on
several fish and mammal species (Aguilar and Borrell 1995;
Bruhn et al. 1995; Brouwer et al. 1999; Gregory and Cyr 2002;
Wright and Welbourn 2002).
Many pathological lesions have previously been reported
in marine mammal species that have been exposed to con-
taminants with known endocrine-disruptive properties
(Gregory and Cyr 2002). Beside reproduction effects, severe
adrenocortical hyperplasia, thymic atrophy, splenic deple-
tion, osteoporosis, intestinal ulcers, claw malformations,
arteriosclerosis, uterine cell tumors, and decreased epidermal
thickness have been reported in grey and ringed seals from
the polluted Baltic Sea (Bergman 1999; Bergman and Ols-
son 1985; Bergman et al. 2002). Few environmental and
laboratory researchers have focused on the disruption of the
thyroid system, despite the fact that many wildlife species in
the world suffer unusual thyroid gland development and
ratios of circulating thyroid hormones (THs) (Colborn 2002).
Thyroid hormones are important regulators of intermediary
metabolism, growth, neural development, and aspects of
reproduction (Gregory and Cyr 2002; Sher et al. 1998). In
Correspondence to: K. Das; email: [email protected]
Arch. Environ. Contam. Toxicol. 51, 720–729 (2006)
DOI: 10.1007/s00244-005-0098-4

cetaceans and other mammals, these hormones are also be-
lieved to play an important role in controlling heat loss
(McNabb 1992).
Organohalogen compounds such as PCBs and PBDE interact
with thyroid hormone expression principally triiodothyronine
T3 and thyroxine T4 in fish, rats, and seals (Brouwer et al.
1989, 1999; Braathen et al. 2004; Hall et al. 1998, 2003;
Woldstad and Jensen 1999; Zhou et al. 2000; Debier et al.
2005). Because of structural similarities, PCBs congeners (and
their hydroxylated metabolites) and thyroid hormones compete
for the same binding sites on transport proteins such as trans-
thyretin (TTR) (Brouwer et al. 1989, 1998). Impairment of
thyroid functions such as a decrease of thyroid levels, colloid
depletion of thyroid follicles, and fibrosis were previously
associated with PCB contamination in harbour seals (Brouwer
et al. 1989; Schumacher et al. 1993). Evaluation of the thyroid
gland morphology may, therefore, be a suitable indicator for
compounds that interfere in thyroid hormone metabolism
(Brouwer et al. 1999; Schumacher et al. 1993).
The level of organohalogen compounds in the blubber of
marine mammals depends upon the species, metabolism, body
condition, sex, age, diet, and geographical area (Jepson et al.
1999, 2005; O0Hara and Becker 2003; O0Shea and Tanabe
2003). High levels of PBDE, DDT, and PCBs were previously
described in porpoises from the North and Baltic Seas com-
pared to northern coasts of Norway, Iceland, and Greenland
(Bruhn et al. 1999; Kleivane et al. 1995; Siebert et al. 2002;
Thron et al. 2004). The question arises about the potential
endocrine disrupter effect in these porpoises encountering high
levels of organic contaminants.
To investigate possible associations between chronic expo-
sition to organic pollutants and thyroid function in harbour
porpoises, the structure of thyroid glands of 57 harbour por-
poises from German and Danish, Norwegian, and Icelandic
waters was examined. The different compartments of the
thyroid gland (follicle, connective, and solid tissues) were
quantified by histomorphometric and immunohistological
analysis and their relative proportion was related to previously
described toxaphene, PCB, DDT, DDE, and PBDE compounds
(Bruhn et al. 1999; Thron et al. 2004; Siebert et al. 2002).
Materials and Methods
Tissue Sampling
Between 1998 and 2001, tissue samples (thyroid, blubber) were col-
lected on 17 porpoises from the German and Danish Baltic Sea, 14
from the German North Sea, 12 from Icelandic, and 14 from Nor-
wegian waters (Fig. 1). Post-mortem investigations were performed
according to standard protocol (Siebert et al. 2001). Of the 57
investigated animals, 19 were males and 38 females. Forty-five har-
bour porpoises were caught animals and 12 porpoises were stranded.
The age was determined by counting the dental growth layers
(Lockyer 1995).
Histology and Immunohistochemistry
Samples of the thyroid glands were fixed in 10%formalin and
embedded in paraffin wax at 60C for histological and immunohis-
tological investigations. Paraffin wax-embedded tissue sections
(5 lm) were stained with hematoxylin and eosin (HE) and by elastic
van Gieson for the detection of collagen.
For immunohistochemistry, a polyclonal rabbit anti-human thyro-
globulin antibody (Code No A 0251, DAKO Corporation) and the
Avidin-Biotin-Peroxidase complex method was used as described
previously (Baumgrtner et al. 1989). The serum used was from a
goat (PAA Laboratories GmbH). A monoclonal antibody against the
T-cell surface antigen of chicken lymphocytes (T1) was used as a
control. The polyclonal antibody against thyroglobulin was used in a
solution of 1:2600 (in TBSc). A biotinylated anti-rabbit-immuno-
globulin (Vector Laboratories Inc., BA 1000) was used as a secondary
antibody. The sections were then treated with avidin-biotin-peroxi-
dase complex (ABC) (Vector Laboratories Inc., PK 4000). As a
positive control, thyroid gland sections were treated with the control
antibody. Furthermore, previously positively stained sections were
used as a control.
Scoring of the Thyroid Gland
For semiquantitative evaluation, ten randomly selected visual fields in
the microscope with a magnification of 200 of each section were
observed. The tissue proportion of the thyroid gland was determined.
The proportion of the follicle, solid, and elastic connective tissue was
estimated and divided into 5 categories (0–5%, 5–25%, 25–50%,
50–75%, 75–100%). For each tissue type, a mean value was assigned
(Table 1). The calculation of the median used for statistical analyses
was performed from 10 values per individual.
Diameter of Thyroid Gland Follicles
Size and quantity of the thyroid gland follicles were registered. Five
images, at microscope magnification of 48 (ocular), were taken of
each thyroid slide (Van Giesson coloration). Furthermore, the number
of follicles with a diameter smaller than 3 mm (small follicles; real
diameter < 62.5 lm), between 3 and 10 mm (medium size follicles;
real diameter: between 62.5 and 208 lm), and larger than 10 mm
(large size follicles, real diameter > 208 lm) were determined.
PCBs, PBDEs, Toxaphene, DDT, and DDE Analyses
The blubber concentrations of 6 individual polychlorinated biphenyl
(PCB) congeners, 5 individual polybrominated diphenylether (PBDE)
congeners, 5 toxaphene congeners, p,p0-dichlorodiphenyltri-
chloroethane (DDT), and p,p0-dichlorodiphenyldichloroethylene
(DDE) were measured using standardized methods (Bruhn et al. 1999;
Thron et al. 2004). Detailed method and results are presented else-
where (Bruhn et al. 1999; Siebert et al. 2002; Thron et al. 2004).
Briefly, after lipid Soxhlet extraction, blubber samples were analyzed
for PCBs, PBDEs, toxaphene, p,p0-DDT, and p,p0-DDE. The levels of 5
individual PBDE congeners (IUPAC nos. 47, 99, 100, 153, 154) were
determined by GC/MS-EI. The levels of 5 individual toxaphene
congeners (Parlar nos. 26, 40, 42, 44, 50) were determined by GC/MS-
NCI. PCB concentrations correspond to the sum of 6 congeners (nos.
99, 149, 138, 153, 180, 187). All data are expressed in ng g
)1
of lipid.
Statistical Analyses
The statistical analysis of the data was performed with the statistical
program package SPSS 11.5.0. Contaminant values were log-trans-
formed to achieve homogeneity of variances and normal distribution.
Thyroid Fibrosis in Harbour Porpoises 721

Thyroid tissue compartments (solid, follicles, and connective tissues)
were standardized such that their sum equaled one. The number of
follicles (small, middle size, and large diameter) were added to a
constant (one) and subsequently log-transformed.
The relationships between thyroid tissue distribution and toxico-
logical data were analyzed in two steps. First, a correlation-based
principal component analysis (PCA) was performed to reduce the 5
original pollutant variables in order to avoid misleading results due to
correlating independent variables (‘‘multi-collinearity’’) in subsequent
analyses. After the transformation, none of the distributions differed
significantly from normality (one-sample Kolmogorov-Smirnov test:
all n‡52, all p‡0.1; no alpha-level adjustment applied). The PCA
was indicated because a considerable proportion of the correlations
between the transformed variables was rather large (seven out of ten
correlation coefficients larger than 0.5) and, correspondingly, justified
by a Kaiser-Meyer-Olkin measure of sampling adequacy equaling
0.606 (which is above the required value of 0.5) as well as BartlettÕs
Test of Sphericity revealing significance (chi-squared = 201,
d.f. = 10, p< 0.001; McGregor 1992). Thereafter, multiple linear
regressions were applied with the two factor scores revealed by the
PCA and age as the independent variables and thyroid tissue pro-
portion (follicles and connective tissues) or numbers of follicles as the
dependent variable. Thus, we calculated a total of five multiple
regressions, one for each of the five dependent variables, separately.
Solid tissue was excluded from the multiple regression because its
proportion exhibited a strong negative correlation with elastic tissue
(Pearson correlation: r
P
=)0.86, n= 58, p< 0.001), meaning that one
of the two variables could explain the other. It must be quoted that this
global approach combining all the studied compounds was preferred
because the pollutants act as a mixture with synergic and antagonistic
behaviour.
The bi-plot (see Fig. 5) is based on PearsonÕs correlation coef-
ficients between percentages of tissues as well as numbers of small,
medium, and large follicles, on the one hand, and principal com-
ponents, on the other hand (Gabriel 1971). In order to reduce the
degree of skewness in the distributions of these variables, we arc-
sine transformed solid and connective tissue percentages and arc-
sine transformed the square root of follicular tissue percentage.
However, these transformations were not sufficient with regard to
revealing normally distributed data. Hence, parametric and non-
parametric correlations were both calculated. However, results of
both tests were the same with regard to the direction and signifi-
cance of correlation coefficients, indicating that conclusions are not
influenced by deviations from the assumptions of parametric cor-
relation. Further on, only parametric correlation coefficients were
reported.
Intersite comparison was investigated using one-way ANCOVA
followed by post-hoc Tukey tests. Since we performed multiple tests
on single null hypotheses (e.g., no difference between the study areas
with regard to tissue variables), an adjustment of the probability of a'-
Fig. 1. Harbour porpoise sampling sites within the northeast Atlantic. G: Germany; D: Denmark; S: Sweden; N: Norway; GB: Great Britain; IR:
Ireland; I: Iceland; NS: North Sea; BS: Baltic Sea
Table 1. Proportion of the solid, follicle, and connective tissue of the
harbour porpoise thyroid observed by microscopy and their related
assigned statistical value (in %)
Proportion of the thyroid tissue Assigned value
0–5 2.5
6–25 15
26–50 37.5
51–75 62.5
722 K. Das et al.

errors was required. p-values for the number of tests of the same
nullhypothesis were thus corrected, applying the equation:
pCk ¼1ð1a0Þk
derived from conversion of the Dunn-idk method (Sokal and
Rohlf 1995) and in which p
Ck
denotes the corrected p-value, a0
equals the originally derived p-value, and kequals the number of
tests. We calculated effect sizes applying the formula indicated in
Bortz (1999).
Results
Histological Investigations
Three tissue compartments were distinguished in the thyroid
gland: epithelium cells, follicles, and connective tissues
(Figs. 2 and 3). Irregular or oval follicular lumens surrounded
by follicular epithelial cells were seen in the parenchyma of
the thyroid (Figs. 2 and 3). Narrow to large connective tissue
strands can be found between the follicles (Fig. 2b). Connec-
tive tissue is typically rich in fibrous tissues and contained
very few cells.
The tissue proportion median in the visual fields in the
microscope ranged from 15 to 87.5%for the solid tissues, 2.5
to 37.5%for the follicles, and 2.5 to 62.5%for the connective
tissues (Table 2).
The rabbit polyclonal antibody designed to react against
human thyroglobulin cross-reacted positively in the thyroid of
the harbour porpoises (Fig. 4). Thyroglobulin was positively
detected in the lumen of the follicles either disseminated in the
parenchyma or in the follicular epithelial cells.
Intersite comparisons revealed that harbour porpoises from
Iceland displayed a lower proportion of connective tissue
compared to porpoises from Germany (North and Baltic Seas)
and Norway (p< 0.001) (Table 3). The proportion of follicles
(%) did not vary between locations (p> 0.1) (one-way
ANCOVA, Table 3). Harbour porpoises from Iceland dis-
played also a higher number of small follicles compared to
porpoises from other locations (p< 0.001).
Fig. 2. Thyroid gland of a harbour porpoise from
German waters. F: follicle. Colloid (C) is
surrounded by thyroid epithelium (between the
arrowheads). S: solid tissue. (A) Hematoxylin and
eosin staining; (B) Van Giesson staining allowing
the identification of the connective tissue (CT) in
red. Magnification: X230
Thyroid Fibrosis in Harbour Porpoises 723

Fig. 3. Thyroid gland of a harbour porpoise from
Icelandic waters. (A) Eosine coloration, C:
colloid. (B) Van Giesson coloration allowing the
identification of the connective tissue (CT) in red.
Magnification: X230
Table 2. Morphometry of the thyroid gland and organochlorine concentrations (ng g
)1
lipid) in the blubber of harbour porpoises from German
and Danish Norwegian, and Icelandic coasts (Siebert et al. 2002; Thron et al. 2004)
German coast
North Sea (n= 14)
German and Danish coasts
Baltic Sea (n= 17)
Icelandic coast
(n= 12)
Norwegian
coast (n= 13)
Tissue proportion
Solid tissue (%) 59 € 14 50 € 15 91 € 0.05 59 € 17
Follicle (%) 8€5 12€15 5€5 6€9
Connective tissue (%) 33 € 15 38 € 16 3 € 1 35 € 15
Number of follicles
<62.5 lm 42 36 353 33
Between 62.5 and 208 lm99 56 65 19
>208 lm2327 216
Pollutants
SPCBs 7664 € 5075 8247 € 7949 1550 € 1517 4710 € 2861
SPBDE 1081 € 1526 219 € 219 52 € 23 481 € 320
SToxaphene 1051 € 1901 445 € 504 1187 € 625 1525 € 493
DDE 1226 € 835 1571 € 1587 896 € 681 1618 € 1122
DDT 255 € 252 428 € 559 226 € 140 674 € 519
Data are given as mean (median) € standard deviation, range of concentrations (minimum-maximum); n: number of samples; nd: not determined.
724 K. Das et al.